COVID-19 detection from CT scans using a two-stage framework.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, is a severe respiratory illness with potentially fatal complications.
• Radiological imaging, including CT scans, plays a crucial role in diagnosing COVID-19.
• Accurate and efficient detection methods are vital for managing the pandemic.

Purpose of the Study:

• To propose an advanced end-to-end framework for automated COVID-19 detection from CT scan images.
• To enhance diagnostic accuracy by integrating deep feature extraction with optimized feature selection.
• To evaluate the framework's performance against existing state-of-the-art methods.

Main Methods:

• Utilized three deep learning Convolutional Neural Networks (CNNs) for deep feature extraction from CT images.
• Implemented a meta-heuristic optimization algorithm, Harmony Search (HS), combined with Adaptive $$ -Hill Climbing (A $$ HC) for feature selection (FS).
• Evaluated the framework on two SARS-CoV-2 CT-Scan Datasets (2482 and 2926 images).

Main Results:

• Achieved high accuracy scores of 97.30% and 98.87% on the respective datasets.
• Outperformed several state-of-the-art optimization algorithms in COVID-19 detection.
• Demonstrated performance comparable to recent advanced methods using the same datasets.

Conclusions:

• The proposed deep learning framework with optimized feature selection offers a highly accurate approach for COVID-19 detection from CT scans.
• This method shows significant potential for improving the speed and reliability of COVID-19 diagnosis.
• The developed feature selection algorithms are publicly available for further research and application.